Basket Trading BALUDA'S CSS Crossovers

[MathiasJ]

...5 Baskets closed now this week but M5H1 is building way too much drawdown so I am going to cut this one now. It confirms the importance of higher timeframe direction. H1D1 drawdown has stayed below 5%.

As they say, the biggest secret to short term trading is to lengthen the time frame (think it was Larry Williams who said that).

I have a question for you TJ, and perhaps Mathias if he'd like to chime in, (some great work you have done with this EA Mathias). Is it possible to counter trend trade with this EA. Example, when the currency line falls below whatever threshold you set in CSSDiff, the EA flags this as a sell, what about if I wanted to counter trade that and make that a buy. I played with the settings by putting -ve numbers in etc. but didn't seem to work.

I'm also having problems with CSSDelta, from my understanding it's the change in the CSS value between whatever 2 bars you set, but the values shown in the dashboard seems to take this difference and x2. The value you are using TJ is also set to 0.04 which seems really quite low to me, that's only a very small change is it not? Maybe I'm not understanding it properly but a higher threshold should lead to better (and less) trades if I'm reading it right. Haven't tested this myself yet, I just want to get a proper understanding of some of the variables.

Hi!
If you want to trade opposite then please set the entry to Operation_Opposite instead of Operation_Normal.
This works for all triggers.

Second question:
It depends on AbsoluteValues true/false. If true then it compare the values against 0 line.
If false then it compare to the other line value and this means value*2 if you have WeighAllSymbols=true.
Try the different combinations to find out how it behaves.
Try WeighAllSymbols true/false and AbsoluteValues true/false combinations to find out
BR
Mathias

[TraderJoeForex]

The value you are using TJ is also set to 0.04 which seems really quite low to me, that's only a very small change is it not? Maybe I'm not understanding it properly but a higher threshold should lead to better (and less) trades if I'm reading it right. Haven't tested this myself yet, I just want to get a proper understanding of some of the variables.

Yes so a delta value of 0.04 is a currency slope that is just slightly +ve in the direction we wish to trade so anything below this is excluded which prevents entering on a retrace until price action is turning around sufficiently for the delta to be + ve (which in real terms is something like price action breaking the high of the previous candle(s) ..depending how many candles you choose to average back). Using 0.04 rather than 0 just adds a small buffer... you may well want to use a higher value to be more selective as you say .. it is a balancing act between this and getting in too late.

[dclayw]

Hi!
If you want to trade opposite then please set the entry to Operation_Opposite instead of Operation_Normal.
This works for all triggers.

Second question:
It depends on AbsoluteValues true/false. If true then it compare the values against 0 line.
If false then it compare to the other line value and this means value*2 if you have WeighAllSymbols=true.
Try the different combinations to find out how it behaves.
Try WeighAllSymbols true/false and AbsoluteValues true/false combinations to find out
BR
Mathias

Thanks Mathias, all sorted now. I was looking at counter trend trading the original CSS a year or 2 back, (counter trading the peaks and troughs on the lower timeframe but in sync with a higher timeframe). Looked promising but I had to quit trading for a while due to unforseen circumstances. Your EA should make it so much easier to test strategies such as this.

[dclayw]

Yes so a delta value of 0.04 is a currency slope that is just slightly +ve in the direction we wish to trade so anything below this is excluded which prevents entering on a retrace until price action is turning around sufficiently for the delta to be + ve (which in real terms is something like price action breaking the high of the previous candle(s) ..depending how many candles you choose to average back). Using 0.04 rather than 0 just adds a small buffer... you may well want to use a higher value to be more selective as you say .. it is a balancing act between this and getting in too late.

Thanks for explanation, and yes I was talking about using a higher value. Visual inspection suggests there is more follow through when the slope of the CSS at entry is steeper, but using a just slightly +ve value makes sense on a lower TF where you do need to get in quick. I'm more interested in the higher TFs (H4 mainly) where it is not so important to get in really early. It's just a matter of getting the time to do the testing as it's the only way to confirm.

[Gertje]

After 19 tradingdays, limiting the EA to 1 basket @ 2% a day, I'm up over 65%

http://www.fxblue.com/users/gertje_demo_1

Using Zennor's setfile, H1 in line with H4 on the CSS.
DD has been high sometimes, with a maximum of 38% on the 14th of Oktober.

[KeithM]

After 19 tradingdays, limiting the EA to 1 basket @ 2% a day, I'm up over 65%

http://www.fxblue.com/users/gertje_demo_1

Using Zennor's setfile, H1 in line with H4 on the CSS.

DD has been high sometimes, with a maximum of 38% on the 14th of Oktober.

Wow. Amazing that it picked itself up from 38%DD. Are you trading all pairs, sessions and what is your position size calculation?

I have also gone back to the original and removed the basket SL. It was a quiet week but did close 2 baskets after some small drawdowns. I have reduced the position size and this keeps DD down but it can wallow for a couple of days before recovering

[TraderJoeForex]

... all good at this end too. I restarted a test mid-week after cutting the M5H1 set as I mentioned previously.

I put a H4D1 set on Live and a H1D1 set on Demo both with GTF and interestingly both delivered fairly similar results over the past few days although for H4D1 it was nothing then everything today whereas H1 delivered at least one basket each day.

This was trying just CSS cross on the lower timeframe rather than Diff and Delta but still keeping Diff and Delta on D1 which seems to have worked well. I only did this to prevent getting 3 trades on one pair from 3 subsequent candles on H1.. the last of which seemed invariably to always end up in drawdown from a few tests I had running on H1 and made the multiple trades per pair slightly counterproductive. Using the cross instead (and level check as > 0.2) but without waiting for a candle close takes longer to build the same size position but with better-averaged entries and has also meant I ended up with a few hedged trades where price changed direction during the day which makes the negative impact of those pairs less within the basket.

I feel I have reached a fairly decent basic safe setup with these sets for smaller account sizes excluding the CHF pairs and filtering by session times and Global Trend Filter. Next I want to look at the range allowed for trade setup on the higher timeframe.

I am wondering whether restricting the D1 diff to a maximum of +4 to prevent entering once a trend is already extended will perhaps protect from getting caught when price turns. Then also maybe allowing a negative value upto -4 to catch reversals in the right direction instead.

Have a great weekend everyone

[Shiva]

After 19 tradingdays, limiting the EA to 1 basket @ 2% a day, I'm up over 65%

http://www.fxblue.com/users/gertje_demo_1

Using Zennor's setfile, H1 in line with H4 on the CSS.
DD has been high sometimes, with a maximum of 38% on the 14th of Oktober.

Brilliant!

[Zennor]

After 19 tradingdays, limiting the EA to 1 basket @ 2% a day, I'm up over 65%

http://www.fxblue.com/users/gertje_demo_1

Using Zennor's setfile, H1 in line with H4 on the CSS.
DD has been high sometimes, with a maximum of 38% on the 14th of Oktober.

Nice one Gertje, the same set with H1 and D1 hit a margin call this week. I think that the D! is just too long a time frame for the hedge trades to be placed in a timely manner to counter the losing trades. :hi:

[Gertje]

Nice one Gertje, the same set with H1 and D1 hit a margin call this week. I think that the D! is just too long a time frame for the hedge trades to be placed in a timely manner to counter the losing trades. :hi:

I've seen margin around 200-300% sometimes, not less.